Data set labeling method, electronic equipment and medium

Abstract

The invention relates to a data set labeling method, electronic equipment and a medium. The method comprises the following steps: S1, acquiring all video segments of a plurality of cameras in a preset time period in a preset geo-fence; S2, extracting a corresponding first to-be-processed image from each video segment; S3, inputting the first to-be-processed image into the RE ID pedestrian recognition model, obtaining RE ID feature vectors, and forming an RE ID feature vector set; S4, marking all the RE ID feature vectors of which the target similarities of the to-be-matched RE ID feature vectors are greater than a preset RE ID feature similarity threshold value and video segments corresponding to the to-be-matched RE ID feature vectors as same unique IDs; and S5, deleting the RE ID feature vector marked with the unique ID, judging whether the current RE ID feature vector set is an empty set or judging whether the similarity of any two RE ID feature vectors is smaller than the RE ID feature similarity threshold, if yes, ending the process, and otherwise, returning to the step S4. According to the method, the data set can be accurately and efficiently labeled.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a data set labeling method, electronic equipment and media. Background technique [0002] The analysis and study of gait is a comprehensive study of human motor function, including the measurement, description and quantitative evaluation of human motion characteristics. Through the analysis and research of gait, the gait cycle can be identified, and the parameters of gait kinematics and dynamics can be calculated. In recent years, the study of gait has played a very important role and application in identification and other aspects. In identification, small changes in gait styles of different people can be used as biometric identifiers to identify individual people. To analyze and study gait, a large amount of gait data is needed. [0003] In the prior art, the video data collected by the camera is usually manually identified to mark the data set, so that the video data corre...

AI Technical Summary

Problems solved by technology

However, due to the large amount of video data and the large number of people involved, it takes a lot of time to manually label the data set, and the labeling efficiency is very low

Images